PROJECT SUMMARY/ABSTRACT I lead the Children?s Oncology Group Phase III clinical trial, ACNS0332, which evaluates treatment options for children with high-risk medulloblastoma (the most common pediatric brain tumor) and supratentorial primitive neuroectodermal tumors (sPNETs). The study opened in 2007 and underwent a major amendment in 2014, when emerging data revealed biological disparity between medulloblastomas and sPNETs as well as heterogeneity in sPNET patients. We discontinued sPNET patient enrollment, and genomic analyses funded by the prior cycle of this grant, revealed that 71% of the non-pineal sPNET patients were actually high grade glioma, ependymoma or atypical teratoid rhabdoid tumors, despite sPNET appearance by histopathology. This reveals the limitations of traditional histopathology and shows that contemporary genomic analyses could spare many children from receiving craniospinal irradiation that is not necessary and not helpful. In Aim 1 of this renewal application, we extend the genomic studies to the 300 medulloblastoma patients in the study. We collected research tissue from over 95% of these patients and anticipate that the studies will reveal 1) patient groups who are likely to die from their disease despite the intense therapy on ACNS0332, 2) patient groups that were placed on ACNS0332 because of clinical or histopathologic observations that may include a mixture of good prognosis patients (e.g., those who would fare well with much less radiation than provided on ACNS0332) as well as those with genomically-predicted poor prognosis, who should be stratified differently in the future. In Aim 2 we address the radiation resistance phenotype of the worst prognosis patients, particularly those with amplified MYC or MYCN. We will collect pre- and post-radiation specimens from patient-derived orthotopic xenograft (PDOX) models (14 MYC/MYCN amplified) that we generated and characterized in the prior cycle of this grant; other PDOX models that we receive from four collaborators; and matching cell lines that we generated and characterized. We will use the cell lines for to screen FDA approved drugs for those that overcome radiation resistance and to conduct functional genomic screens to identify pathways that, when inhibited, convert radiation resistant cells into radiation sensitive cells. In vivo efficacy studies on PDOX mouse models representing dozens of patients will follow. The significance is that this work will likely reduce unnecessary radiation exposure to patients who do not warrant high-dose craniospinal irradiation, identify patients who would best be served by alternative therapies, and generate pre-clinical data to prioritize the most effective agents for upcoming human clinical trials.